Sparsity-based space-time adaptive processing using convolutional neural network

• DOI: 10.1049/icp.2021.0487
• Publication Date: 2021-06-30T20:07:25Z
• Authors: K. Duan; X. Yang; H. Xie; Y. Wang

In this paper, a deep learning framework for space-time adaptive processing is developed. Firstly, set of clutter covariance matrixes (CCMs) are modeled based on the prior parameters radar and navigation system with respect to all possible levels non-ideal factors, columns each CCM formulated as undersampled noisy linear measurements sparse coefficients corresponding angle-Doppler spectrum. Then original spectrum coefficients, obtained by least-square estimation from CCMs known steering dictionary, used input train convolutional neural network (CNN). Meanwhile, labels can be exact via minimum variance distortionless response algorithm. Once trained, CNN predict that corresponds new measurement vector in near real time. Simulations results have demonstrated superiority proposed method both suppression performance computation efficiency.